Robust Hydrolysis of Prostaglandin Glycerol Esters by Human ...

14 downloads 91 Views 1MB Size Report
Oct 22, 2015 - DulbeccoLs modified EagleLs medium; GCLM, glutamate-cysteine ligase modifier subunit; HA, hemagglutinin; HBS, Hepes-buffered saline; ...
Supplemental material to this article can be found at: http://molpharm.aspetjournals.org/content/suppl/2014/08/20/mol.114.094284.DC1.html 1521-0111/86/5/522–535$25.00 MOLECULAR PHARMACOLOGY Copyright ª 2014 by The American Society for Pharmacology and Experimental Therapeutics

http://dx.doi.org/10.1124/mol.114.094284 Mol Pharmacol 86:522–535, November 2014

Robust Hydrolysis of Prostaglandin Glycerol Esters by Human Monoacylglycerol Lipase (MAGL) s Juha R. Savinainen, Emilia Kansanen, Tatu Pantsar, Dina Navia-Paldanius, Teija Parkkari, Marko Lehtonen, Tuomo Laitinen, Tapio Nevalainen, Antti Poso, Anna-Liisa Levonen, and Jarmo T. Laitinen School of Medicine, Institute of Biomedicine (J.R.S., D.N-P., Te.P., J.T.L.), A.I. Virtanen Institute for Molecular Sciences (E.K., A-L.L.), School of Pharmacy (Ta.P., Te.P., M.L., T.L., T.N., A.P.), Faculty of Health Sciences, University of Eastern Finland, Kuopio, Finland

ABSTRACT The primary route of inactivation of the endocannabinoid 2-arachidonoylglycerol in the central nervous system is through enzymatic hydrolysis, mainly carried out by monoacylglycerol lipase (MAGL), along with a small contribution by the a/b-hydrolase domain (ABHD) proteins ABHD6 and ABHD12. Recent methodological progress allowing kinetic monitoring of glycerol liberation has facilitated substrate profiling of the human endocannabinoid hydrolases, and these studies have revealed that the three enzymes have distinct monoacylglycerol substrate and isomer preferences. Here, we have extended this substrate profiling to cover four prostaglandin glycerol esters, namely, 15-deoxy-D12,14-prostaglandin J2-2-glycerol (15d-PGJ2-G), PGD2-G, PGE2-G, and PGF2a-G. We found that the three enzymes hydrolyzed the tested substrates, albeit with distinct rates and preferences. Although human ABHD12 (hABHD12) showed only marginal activity toward PGE2-G, hABHD6

Introduction The endocannabinoids anandamide (N-arachidonoylethanolamide, AEA) and 2-arachidonoylglycerol (2-AG) are v-6 arachidonic acid (AA)-containing lipids involved in a broad range of processes, including synaptic transmission, food intake, energy balance, pain, and inflammation (Di Marzo et al., 2007; Kano et al., 2009; Alhouayek et al., 2014). The endocannabinoids signal via two G protein–coupled receptors (cannabinoid type 1 receptor [CB1R] This research was supported by the Academy of Finland [Grants 139620, 139140, 128056, 133290]; Biocenter Finland/DDCB, the Orion Farmos Foundation, the Emil Aaltonen Foundation, and the Sigrid Juselius Foundation. J.R.S. and E.K. contributed equally to this work. dx.doi.org/10.1124/mol.114.094284. s This article has supplemental material available at molpharm.aspetjournals. org.

preferentially hydrolyzed PGD2-G, and human MAGL (hMAGL) robustly hydrolyzed all four. This was particularly intriguing for MAGL activity toward 15d-PGJ2-G whose hydrolysis rate rivaled that of the best monoacylglycerol substrates. Molecular modeling studies combined with kinetic analysis supported favorable interaction with the hMAGL active site. Long and short MAGL isoforms shared a similar substrate profile, and hMAGL hydrolyzed 15d-PGJ2-G also in living cells. The ability of 15d-PGJ2-G to activate the canonical nuclear factor erythroid 2-related factor (Nrf2) signaling pathway used by 15d-PGJ2 was assessed, and these studies revealed for the first time that 15d-PGJ2 and 15d-PGJ2-G similarly activated Nrf2 signaling as well as transcription of target genes of this pathway. Our study challenges previous claims regarding the ability of MAGL to catalyze PG-G hydrolysis and extend the MAGL substrate profile beyond the classic monoacylglycerols.

and type 2 receptor [CB2R]) that show unique and tissuespecific distribution. Cannabinoid CB1R is widely expressed in the nervous system, mainly in axonal terminals of central and peripheral neurons whereas CB2R is principally expressed in cells of the immune system. Besides engaging G protein– coupled receptors, recent evidence implicates members of the transient receptor potential (TRP) ion channel family as ionotropic cannabinoid receptors (Caterina 2014). The signaling function of anandamide is primarily terminated by fatty acid amide hydrolase (FAAH) with AA and ethanolamine formed as the products. The major route for 2-AG inactivation is via enzymatic hydrolysis, liberating AA and glycerol. At the bulk brain level in vitro, MAGL accounts for ∼85% of this hydrolysis, and the remaining ∼15% is chiefly attributed to ABHD6 and ABHD12, two recently discovered

ABBREVIATIONS: AA, arachidonic acid; ABHD, a/b-hydrolase domain; ABPP, activity-based protein profiling; 1-AG, 1-arachidonoylglycerol; 2-AG, 2-arachidonoylglycerol; BSA, bovine serum albumin; CB, cannabinoid; COX-2, cyclooxygenase 2; 1-DG, 1-decanoyl-rac-glycerol; DMEM, Dulbecco’s modified Eagle’s medium; GCLM, glutamate-cysteine ligase modifier subunit; HA, hemagglutinin; HBS, Hepes-buffered saline; HEK, human embryonic kidney; HMOX1, heme oxygenase 1; HPLC, high-performance liquid chromatography; HUVEC, human embryonic umbilical vein endothelial cells; JJKK-048, {4-[bis-(benzo[d][1,3]dioxol-5-yl)methyl]-piperidin-1-yl}(1H-1,2,4-triazol-1-yl) methanone; MAG, monoacylglycerol; MAGL, monoacylglycerol lipase (monoglyceride lipase); NQO1, NAD(P)H quinone oxidoreductase-1; Nrf2, nuclear factor erythroid 2-related factor; PGD2-G, prostaglandin D2-1-glycerol ester; PGE2-G, prostaglandin E2-1-glycerol ester; PGF2a-G, prostaglandin PGF2a-1-glycerol ester; 15d-PGJ2-G, 15-deoxy-D12,14-prostaglandin J2-2-glycerol ester; PPARg, peroxisome proliferator-activated receptor-g; TAMRA-FP, tetramethylrhodamine fluorophosphonate. 522

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015

Received June 25, 2014; accepted August 19, 2014

Prostaglandin Glycerol Esters as In Vitro Substrates of MAGL

Materials and Methods Drugs, Chemicals, and Reagents. All components of the glycerol assay mix (Navia-Paldanius et al., 2012), as well as 1-DG (1-decanoyl-rac-glycerol) were purchased from Sigma-Aldrich (St. Louis, MO). We synthesized 1(3)-AG in our laboratory or bought it from Cayman Chemicals (Ann Arbor, MI). We purchased 2-AG, 15-deoxy-D12,14-prostaglandin J2 (15d-PGJ2), 15-deoxy-D12,14-prostaglandin J2-2-glycerol (15d-PGJ2-G); PGD2-G; PGE2-G, and PGF2a-G from Cayman Chemicals. The chemical structures and calculated logP values of the PG-G species are shown in Supplemental Fig. 1. While 15d-PGJ2-G represented the 2-isomer, PGD2-G, PGE2-G, and PGF2a-G were 1(3)-isomers. The MAGL inhibitor JJKK-048 {4-[bis-(benzo[d][1,3]dioxol-5-yl)methyl]-piperidin-1-yl} (1H-1,2,4-triazol-1-yl)methanone) was synthesized as previously described elsewhere (Aaltonen et al., 2013). Human Embryonic Kidney 293 Cells with Transient Overexpression of Serine Hydrolases. Our procedures have been described elsewhere (Navia-Paldanius et al., 2012); briefly, human embryonic kidney 293 (HEK293) cells were cultured as monolayers in Dulbecco’s modified Eagle’s medium (DMEM) (Euroclone, Milan, Italy) containing 10% fetal bovine serum (Euroclone) under antibiotics (penicillin/streptomycin; Euroclone) at 37°C in a humidified atmosphere of 5% CO2/95% air. All plasmids were purchased from Origene Technologies (Rockville, MD). Plasmid for the short isoform of hMAGL was generated by point-mutating the start codon (Met1Leu) of the long isoform by using the Stratagene QuikChange Site-Directed Mutagenesis Kit (La Jolla, CA). Plasmids were introduced to cells by a standard transient transfection procedure using the X-tremeGENE Hp DNA Transfection reagent (Roche, Mannheim, Germany). HEK293 and/or Mock cells (cells transfected with an empty vector) were cultured in parallel for controlling expression and activity in later experiments. Cellular lysates were prepared and protein concentrations measured as previously described elsewhere (Navia-Paldanius et al., 2012). In studies with intact cells, a stable hMAGL-transfected HEK239 cell line was used. The generation and culturing of this line has been previously described elsewhere (Savinainen et al., 2010). Human Embryonic Umbilical Vein Endothelial Cells. Human embryonic umbilical vein endothelial cells (HUVEC) were isolated from umbilical cords (unspecified sex) obtained from the maternity ward of the Kuopio University Hospital, after approval by the Kuopio University Hospital ethics committee. Cells were cultivated as previously described elsewhere (Levonen et al., 2001). Hydrolase Assays. Hydrolase activity assays for hABHD6, hABHD12, and hMAGL, based on fluorescent glycerol detection were conducted in a 96-well-plate format as previously described elsewhere (Navia-Paldanius et al., 2012). The amount of lysate protein per well was 0.3 or 1 mg for HEK293 cells and HUVEC, respectively. The substrate preferences were determined in TEMN-BSA buffer [50 mM Tris-HCl pH 7.40, 1 mM EDTA, 5 mM MgCl2, 100 mM NaCl, supplemented with 0.5% (w/v) fatty acid free bovine serum albumin and the test substrates [5 or 25 mM final concentration, or 3–50 mM in assays for Km determination)]. For each tested substrate, assay blanks without enzyme, cellular background (HEK293/Mock cell lysates), and a glycerol standard was included to monitor assay performance. Fluorescence of the assay blank was subtracted from all values before calculation of the final results based on the glycerol standard. Assessment of PG-G Hydrolysis in Intact HEK293 Cells. HEK293 cells with or without stable overexpression of 3x hemagglutinin (3HA)-tagged hMAGL (Savinainen et al., 2010) were seeded on poly-L-lysine-coated wells of a 96-well plate and cultured in DMEM (200 ml/well) containing 10% fetal bovine serum, antibiotics, and additionally the selection antibiotic G-418 (100 mg/ml) as described earlier. After 2 days, DMEM was replaced with Hepes-buffered saline (HBS; 20 mM Hepes, pH 7.40, 140 mM NaCl, 2.5 mM MgSO4, 1 mM KCl, 1 mM CaC2) supplemented with 0.2% (w/v) BSA and 10 mM D(1)-glucose. Cells were pretreated for 1 hour with solvent (0.1% DMSO) or the MAGL-selective inhibitor JJKK-048 (1027 or 1026 M final concentration) in a volume of 100 ml, after which fresh HBS containing either

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015

a/b-hydrolase domain (ABHD) containing serine hydrolases (Blankman et al., 2007; Savinainen et al., 2012). Recent studies have indicated that in certain tissues such as the brain MAGL acts as a key metabolic node capable of linking together the endocannabinoid and the prostaglandin (PG) lipid signaling pathways (Mulvihill and Nomura 2013). This is because MAGL-catalyzed 2-AG hydrolysis generates AA, the key substrate for cyclooxygenase-2 (COX-2)–dependent oxygenation to produce eicosanoids such as prostaglandin E2 (PGE2) and prostaglandin D2 (PGD2). In animal models of neurodegenerative diseases, MAGL-driven 2-AG hydrolysis has been shown to feed the principal substrate for COX-2–derived neuroinflammatory prostaglandins. Moreover, both genetic and pharmacologic MAGL inhibition has been found to exert antiinflammatory and neuroprotective effects (Nomura et al., 2011; Chen et al., 2012; Piro et al., 2012). Thus, MAGL inhibitors may have therapeutic potential. Comprehensive MAGL inactivation drastically alters brain lipid composition, most notably elevating levels of 2-AG and concomitantly decreasing those of AA. The resulting 2-AG overflow brings about functional antagonism of the endocannabinoid system (Chanda et al., 2010; Schlosburg et al., 2010). Under such conditions, 2-AG spillover might be increasingly canalized toward alternative metabolic pathways such as oxidation, catalyzed by COX, lipoxygenases (LOX), or cytochrome P450 family members (Rouzer and Marnett 2011). In particular, COX-2–mediated oxygenation is emerging as the third major metabolic pathway for endocannabinoids with potential therapeutic implications (Hermanson et al., 2014). COX-2–catalyzed 2-AG oxidation is thought to generate prostaglandin glycerol esters (PG-G) via pathways closely matching those of the respective prostaglandins formed from AA (Rouzer and Marnett 2011; Hermanson et al., 2014). Indeed, recent studies have demonstrated that under specific conditions, 2-AG–derived PG-G species are formed by COX-2 action in vitro and in vivo and that these metabolites likely have bioactivities of their own, mediated via as-yet-unidentified nonPG receptors (Rouzer and Marnett 2011; Hermanson et al., 2014). In this scenario, the lifetime and signaling function of the PG-G species would be expected to be under tight enzymatic control, yet very little is known on the metabolic fates of PG-G in vivo or in vitro. The prevailing view is that rapid hydrolysis of PG-G is performed by unknown enzymatic activity, distinct from the endocannabinoid hydrolases fatty acid amide hydrolase (FAAH) and MAGL (Rouzer and Marnett, 2011). This view relies on the findings of a single study reporting that PGE2-G and PGF2a-G were poor substrates for the purified rat MAGL (Vila et al., 2007). However, our recent studies and those conducted in other laboratories are questioning the validity of this view by demonstrating that human MAGL efficiently hydrolyzed various PG-G species (Xie et al., 2010; Laitinen et al., 2014). While extending the substrate repertoire to cover glycerolipid substrates beyond the classic monoacylglycerol (MAG) species, we observed that human endocannabinoid hydrolases readily used PG-G species, albeit with distinct preferences and catalytic rates. This was particularly evident for human MAGL (hMAGL), for which 15d-PGJ2-G was found to be among the best in vitro substrates identified to date. The initial observations were recently reported as part of a study characterizing the role of conserved cysteine residues in the MAGL substrate preference and inhibitor profile (Laitinen et al., 2014). Here, we report our findings regarding hMAGL-catalyzed hydrolysis of PG-G species in more detail.

523

524

Savinainen et al. pGL3-SV40 as control or pGL3-SV40-2xGCLM-ARE-luciferase (Hurttila et al., 2008), and for normalization, with 20 ng of pCMV-b-gal vector. At 24 hours after transfection, the cells were treated with the indicated concentrations of 15d-PGJ2 or 15d-PGJ2-G in DMEM supplemented with 0.5% (w/v) BSA. At 16 hours after treatment, luciferase activities were measured using the Britelite Reporter Gene Assay (PerkinElmer Life Sciences, Waltham, MA) and Victor 1420 Multilabel Counter (Wallac, Turku, Finland). Luciferase activities were normalized to b-galactosidase activities measured as previously described elsewhere (Hurttila et al., 2008). The results are presented as fold-change versus pGL3-SV40-control vector for each treatment. RNA Isolation and Quantitative Real-Time Polymerase Chain Reaction. HEK293 cells with or without (mock) stable 3HA-hMAGL expression were seeded on 12-well plates and treated with 15d-PGJ2 or 15d-PGJ2-G for 8 hours. The cells were collected, and the RNA was extracted with TRI Reagent (Sigma-Aldrich). The cDNA was produced with the Transcriptor First Strand Synthesis Kit (Roche, Mannheim, Germany). The relative expression levels were measured according to the manufacturer’s protocol with quantitative real-time polymerase chain reaction (StepOnePlus Real-Time PCR systems; Applied Biosystems, Foster City, CA) using specific assayson-demand (Applied Biosystems) target mixes for heme oxygenase 1 (HMOX1, Hs00157965_m1), glutamate-cysteine ligase modifier subunit (GCLM, Hs00157694_m1), and NAD(P)H quinone oxidoreductase-1 (NQO1, Hs00168547_m1). The expression levels were normalized to b2-microglobulin (Hs00187842_m1) expression and are presented as fold change in the expression versus control. Molecular Modeling. Molecular modeling was performed using the Schrödinger Maestro software package (Schrödinger Suite 2013-3; Schrödinger, LLC, New York, NY). Structures of small molecules were prepared using the LigPrep module. X-ray crystal structure of the human MAGL was used (pdb:3PE6) (Schalk-Hihi et al., 2011), and the biomolecule was preprocessed using the protein preparation wizard to optimize the hydrogen bonding network and to remove any possible crystallographic artifacts (Protein Preparation Wizard; Schrödinger, LLC). Molecular docking studies were performed following the Induced Fit docking protocol. The grid box was centered using the corresponding X-ray ligand as the template, and at least one hydrogen bond was required to be formed with Ala51 or Met123. Initial docking poses were generated using Glide, after which residues close to the ligand pose (5Å) were further refined using Prime. Redocking was performed using the SP settings of Glide. The graphical illustrations were generated using MOE software (Molecular Operating Environment, version 2013.08; Chemical Computing Group, Montreal, QC, Canada). Data Reproducibility and Statistical Analyses. All numerical data are presented as mean 6 S.E.M. or S.D. from at least two independent experiments, as detailed in Results. The Km and Vmax values were calculated from nonlinear regressions using GraphPad Prism 5.0 for Windows (GraphPad Software, San Diego, CA). Statistical comparisons were performed using one-way analysis of variance, followed by Tukey’s nonparametric test. Asterisks denote statistical significance (*P , 0.05, **P , 0.01, and ***P , 0.001).

Results PG-G Hydrolysis by the Human Endocannabinoid Hydrolases; hABHD6 Readily Uses PGD2-G, and hMAGL Shows Highest Preference Toward 15d-PGJ2-G. Our previous study disclosed distinct MAG substrate profiles for the human serine hydrolases ABHD6, ABHD12, and MAGL (Navia-Paldanius et al., 2012). Here, we have extended this comparative substrate profiling to cover COX-2–derived oxygenation products of 2-AG by testing the utilization of four commercially available PG-G species: 15d-PGJ2-G, PGD2-G, PGE2-G, and PGF2a-G, at 25 mM concentration. The chemical

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015

solvent (0.4% ethanol) or 15d-PGJ2-G (25 mM final concentration) was added in a volume of 150 ml, followed by 120 minutes of incubation at 37°C. Aliquots (40 ml) of buffer were drawn at time points 0 (immediately after gentle plate shaking), 30, 60, and 120 minutes after substrate addition. The glycerol content was determined based on a glycerol standard. Verification of hMAGL-Catalyzed 15d-PGJ2-G Hydrolysis by High-Performance Liquid Chromatography. Incubations with assay conditions mimicking the glycerol assays were performed for 90 minutes using lysates of Mock-HEK293 or hMAGL-HEK293 cells. The incubation volume was scaled up to 400 ml. At time points of 0 and 90 minutes, 100-ml aliquots were removed, acetonitrile (200 ml) was added to stop the enzymatic reaction, and the pH of the samples was simultaneously decreased to 3.0 with phosphoric acid (H3PO4, added to acetonitrile) to stabilize the 2-isomer against acyl migration (Saario et al., 2004). Before the high-performance liquid chromatography (HPLC) analysis, samples were centrifuged at 14,000 rpm for 4 minutes (room temperature), 150 ml of the supernatant was taken into a HPLC vial, and 50 ml of 30 mM H3PO4 was added. The samples were analyzed by the Agilent 1100 HPLC system (Agilent Technologies, Waldbronn, Karlsruhe, Germany), which consisted of a binary pump, a vacuum degasser, autosampler, and a UV detector. The separations were accomplished by a Zorbax Eclipse XDB-C18 reversed-phase column (4.6  50 mm, 1.8 mm; Agilent Technologies, Palo Alto, CA). The UV detector was set at 310 nm. The injection volume was 10 ml. An isocratic-mobile phase mixture of 45% phosphate buffer (30 mM, pH 3.0) in acetonitrile was used at a flow rate of 2.0 ml/min21 at 50°C. Retention times were 2.3 minutes and 2.5 minutes for the 2- and 1 (3)-isomers of 15d-PGJ2-G, respectively, and 4.2 minutes for 15d-PGJ2. The external standard method was used to calculate the concentration of 15d-PGJ2. Activity-Based Protein Profiling. Activity-based protein profiling (ABPP) was conducted using the serine hydrolase-targeting fluorophosphonate (FP) probe tetramethylrhodamine (TAMRA-FP) (ActivX Fluorophosphonate Probes, Thermo Fisher Scientific, Rockford, IL) as previously described elsewhere (Navia-Paldanius et al., 2012). Briefly, lysates (25 or 100 mg protein for HEK293 cells or HUVEC, respectively) were preincubated with 0.5 ml of vehicle (DMSO) or JJKK-048 (1027 M final concentration) for 1 hour at room temperature. After this, the serine hydrolases were labeled with TAMRA-FP (2 mM final concentration) for 1 hour at room temperature. The reaction was stopped by adding a 2x SDS-loading buffer, and the proteins resolved in SDS-PAGE together with molecular weight standards. Protein bands were visualized after in-gel fluorescent gel scanning using a Fujifilm (FLA-3000; Tokyo, Japan) scanner (ex. 532 nm/em. 580 nm). Detection of Nuclear and Cytoplasmic Nuclear Factor Erythroid 2-Related Factor. Immunoblotting against nuclear and cytoplasmic nuclear factor erythroid 2-related factor (Nrf2) was performed essentially as previously described (Jyrkkänen et al., 2008). Briefly, after 2 hours of treatment, the cells were lysed, and the total protein concentration was measured with the BCA-assay (Pierce Biotechnology, Rockford, IL). Nuclear and cytoplasmic fractions were isolated using NE-PER Nuclear and Cytoplasmic Extraction Reagents Kit (Pierce Biotechnology). We used 10 to 20 mg of protein for electrophoresis. The proteins were transferred to nitrocellulose membrane, blocked, and incubated with primary antibodies. The primary antibodies used were rabbit polyclonal anti-Nrf2 (Santa Cruz Biotechnology, Santa Cruz, CA) and rabbit polyclonal anti-Lamin B1 (Abcam, Cambridge, United Kingdom). Blots were visualized using horse radish peroxidase–conjugated secondary antibodies and Supersignal chemiluminescence substrate (Pierce Biotechnology). Blots were visualized using ECL Plus Western Blotting Detection System with Typhoon 9400 (GE Healthcare, Little Chalfont, United Kingdom). Protein expression data were quantified with ImageQuant TL 7.0 software (GE Healthcare). Nrf2 Reporter Gene Assay. Mock-HEK293 and HEK293 cells stably expressing 3HA-hMAGL (Savinainen et al., 2010) were seeded on 96-well plates and transfected the next day with the calcium phosphate transfection method using the following plasmids: 20 ng of

Prostaglandin Glycerol Esters as In Vitro Substrates of MAGL

PGD2-G at rates corresponding to ∼60% of those for 1-AG. In contrast, the hydrolysis rates of the other PG-G species were modest, being ∼15–30% of those measured for 1-AG. For hABHD6, the relative hydrolysis rate was 1-AG . PGD2-G . 15d-PGJ2-G . PGF2a-G @ PGE2-G. For hABHD12, the preferred MAG substrate was 1(3)-AG, and it sparingly used other tested MAG species (Navia-Paldanius et al., 2012). Here, we found that the hABHD12-dependent hydrolysis rates of the PG-G species were modest, being merely ∼10–20% of those determined for 1-AG (Fig. 1C). For hABHD12, the relative hydrolysis rate was 1-AG . . PGE2-G . PGD2-G @ PGF2a-G @ 15d-PGJ2-G. As previously demonstrated (Navia-Paldanius et al., 2012), the high sensitivity of the fluorescent glycerol assay afforded hydrolase activity measurements with only 0.3 mg of lysate protein per well, ensuring that background cellular activity was negligible also with the current set of substrates (Fig. 1D). As PG-G hydrolysis was particularly robust for MAGL, we focused rest of these studies to the characterization of this response in more detail. Verification of hMAGL-Catalyzed 15d-PGJ2-G Hydrolysis by HPLC. Hydrolysis of 15d-PGJ2-G yields equimolar quantities of 15d-PGJ2 and glycerol as the end products (Fig. 2A). To complement the results of the glycerol assay, a HPLC-based method to detect 15d-PGJ2 was employed. The HPLC analysis revealed that the substrate eluted as two peaks: the major peak (∼90%) representing the 2-isomer and the minor peak (∼10%), representing the 1(3)-isomer (Fig. 2B). Given that nonenzymatic acyl migration is expected to take place in aqueous milieu (Saario et al., 2004), incubating 15d-PGJ2-G together with

Fig. 1. PG-G hydrolysis rates in relation to 1-AG hydrolysis by the human endocannabinoid hydrolases. Lysates prepared from HEK293 cells transiently overexpressing hMAGL, hABHD6, or hABHD12 were used as enzymatic source in kinetic fluorescent glycerol assays as previously described (Navia-Paldanius et al., 2012). Cellular lysates (0.3 mg/well) were incubated together with the indicated PG-G species or 1-AG [25 mM final concentration, added from 10 mM stock solutions in ethanol into the glycerol assay mix containing 0.5% (w/v) BSA and 1% (v/v ethanol)]. To stay in the linear range of the standards, glycerol production was monitored kinetically for 30 minutes (A) or 90 minutes (B–D). (A) PG-G substrate profile of hMAGL. (B) PG-G substrate profile of hABHD6. (C) PG-G substrate profile of hABHD12. (D) Cellular background activity for the tested substrates. For comparative purposes, hydrolysis rates of 1-AG are shown for each hydrolase. Data are mean 6 S.E.M. from two to three independent experiments.

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015

structures and calculated logP values of the PG-G species are shown in Supplemental Fig. 1. While 15d-PGJ2-G represented the 2-isomer, the other PG-G species were 1(3)-isomers. For reference purposes, we also tested in parallel 1(3)-AG, as this is the preferred endocannabinoid isomer for hABHD6 and hABHD12, whereas hMAGL uses both isomers at similar rates (Navia-Paldanius et al., 2012; Laitinen et al., 2014). To facilitate comparison with the previous MAG profiling data, we used the TEMN buffer ([50 mM Tris-HCl pH 7.40, 1 mM EDTA, 5 mM MgCl2, 100 mM NaCl], pH 7.4) supplemented with 0.5% BSA (w/v), as this buffer system has proven optimal for studies exploring 2-AG–evoked cannabinoid receptor signaling and endocannabinoid degradation in cellular lysates and native membrane preparations (Savinainen et al., 2003; Saario et al., 2004; Navia-Paldanius et al., 2012; Laitinen et al., 2014). In accordance with our preliminary findings with the ability of 3HA-hMAGL (short isoform) to hydrolyze PG-G species (Laitinen et al., 2014), untagged hMAGL (long isoform) readily hydrolyzed all tested PG-G species, and notably at 2- to 3-fold higher rates as compared with 1-AG (Fig. 1A). The best substrate in this series was 15d-PGJ2-G. Also, it is notable that the hydrolysis rates of 15d-PGJ2-G rivaled those of the best MAG species, previously determined under identical assay conditions (Navia-Paldanius et al., 2012). For hMAGL, the relative hydrolysis rate for the tested substrates was 15d-PGJ2-G . PGD2-G . PGF2a-G @ PGE2-G . 1-AG. Although the MAG substrate profiles of hMAGL and hABHD6 were previously shown to be closely related (Navia-Paldanius et al., 2012), the utilization of PG-G species was a clear deviation from this. As shown in Fig. 1B, hABHD6 readily hydrolyzed

525

526

Savinainen et al.

Mock-HEK lysates for 90 minutes resulted in a notable change in the 1(3)-to-2-isomer ratio, and importantly, no detectable amount of 15d-PGJ2 was formed (Fig. 2C). In parallel incubations together with hMAGL-HEK lysates, a clear peak was observed, eluting at the position of 15d-PGJ2 standard, and with

a concomitant decrease in the peak corresponding to 15d-PGJ2-G (Fig. 2D). Quantification based on the peak area corresponding to 15d-PGJ2 from three independent experiments indicated that while 15d-PGJ2 was undetectable in all experiments with Mock-HEK lysates, parallel incubations with hMAGL-HEK

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015

Fig. 2. HPLC verification of hMAGL-catalyzed 15d-PGJ2-G hydrolysis. (A) Reaction scheme for 15d-PGJ2-G hydrolysis generating equimolar quantities of 15d-PGJ2 and glycerol. (B) Chromatographs showing the elution pattern of 15d-PGJ2-G and 15d-PGJ2 (both from Cayman Chemicals). Note that the substrate is a mixture of the 2- and 1(3)isomers (∼90 and ∼10%, respectively). (C) HPLC analysis of the 15d-PGJ 2 -G metabolites after incubation for 90 minutes with Mock-HEK293 lysates. Note the change in the ratio of the substrate isomers due to nonenzymatic acyl migration in an aqueous milieu. Note also the absence of any product eluting at the position of 15dPGJ2. (D) HPLC analysis of the 15d-PGJ2-G metabolites after incubation for 90 minutes with hMAGL-HEK293 lysates. Note the presence of a new peak eluting at the position of 15d-PGJ2. Data are from one experiment that was repeated twice. Quantification based on the peak area corresponding to 15d-PGJ2 from three independent experiments indicated that although 15d-PGJ2 was below the detection limit in incubations using MockHEK293 lysates, parallel incubations with hMAGLHEK293 lysates yielded 1.5 6 0.1 nmol of 15d-PGJ2 (mean 6 S.E.M., n = 3).

Prostaglandin Glycerol Esters as In Vitro Substrates of MAGL

previously reported (Navia-Paldanius et al., 2012), hMAGL-long migrated mainly as a ∼33–35 kDa doublet (Fig. 3A), whereas the short isoform migrated as a single band of ∼33 kDa (human) or ∼34 kDa (mouse). The crude substrate profiling of MAGL isoforms was performed using 2-AG, 15d-PGJ2-G and the preferred MAG substrate 1-decanoyl-rac-glycerol (1-DG) (NaviaPaldanius et al., 2012), all tested at 5 and 25 mM final concentrations. These experiments revealed that the MAGL splice variants shared similar substrate profile in that 1-DG was used at the highest rates, and all isoforms showed a clear preference for 15d-PGJ2-G over 2-AG (Fig. 3, B–D). Substrate Availability for hMAGL-Catalyzed Hydrolysis in Assays with or without BSA. The experiments described so far were all conducted using our routine assay protocol with BSA (0.5% w/v) included as a carrier for liphophilic compounds. Using the routine protocol, we determined the Km and Vmax values for hMAGL (long isoform) toward 15d-PGJ2-G covering the concentration range 3–50 mM. Due to robust expression of the hMAGL-gene construct in HEK293 cells, the lysates were tested at two dilutions (1:10 and 1:20, both containing 0.3 mg protein/well, diluted with HEK lysates) to ensure that substrate availability was not a limiting factor and that the activity measurements were within the linear range of the glycerol standards (Navia-Paldanius et al., 2012). Nonlinear fitting of the substrate-velocity data indicated that the Km

Fig. 3. Substrate profiles of long and short MAGL isoforms. (A) Activity-based protein profiling with the catalytic serine targeting probe TAMRA-FP of HEK293 cell lysates after transient expression of human MAGL-long (313 aa), human MAGL-short (303 aa), or mouse MAGL-short (303 aa) isoforms. After resolving on SDS-PAGE, serine hydrolases were visualized by in-gel fluorescence scanning. Molecular weight markers (kDa) are indicated on the right. Lysates were pretreated either with dimethylsulfoxide or the MAGL-selective inhibitor JJKK-048 (1027 M) to demonstrate that TAMRA-FP labels catalytically active MAGL isoforms. Note that hMAGL-long migrates principally as a ∼33–35 kDa doublet whereas the short isoforms migrate principally as single bands (hMAGL-short ∼33 kDa, mMAGL-short ∼34 kDa). (B–D) Lysates (0.3 mg/well) of HEK293 cells expressing different MAGL isoforms were incubated together with the indicated substrates [2-AG, 15d-PGJ2-G, or 1-decanoyl-rac-glycerol (1-DG), 5 or 25 mM final concentration, added from 2 or 10 mM stock solutions in ethanol into the glycerol assay mix containing 0.5% (w/v) BSA and 1% (v/v ethanol)]. Glycerol output was determined at a time point of 30 minutes where the substrate availability was not rate limiting. Data are mean + S.D. of duplicate wells from a representative experiment that was repeated once with similar outcome. Statistical comparison was performed using one-way analysis of variance, followed by Tukey’s multiple comparisons; statistical significance is indicated with asterisks (***P , 0.001).

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015

lysates yielded 1.5 6 0.1 nmol of 15d-PGJ2 (mean 6 S.E.M., n 5 3), a value closely matching the estimated amount of glycerol formed at the time point 90 minutes, extrapolated from the reaction rate at time points 0–30 minutes of the kinetic fluorescent assay. Thus, the HPLC analysis validated results of the glycerol assay showing that MAGL-catalyzed 15d-PGJ2-G hydrolysis generates the two expected end products. MAGL Splice Variants Use 15d-PGJ2-G at Similar Rates. Based on molecular and Western blotting evidence, MAGL is expressed in a tissue-specific manner, usually migrating as a single ∼33 kDa protein band or as a ∼33–35 kDa doublet, depending on the tissue (Karlsson et al., 2001). Two MAGL isoforms are reported in the GenBank (long isoform 1; 313 amino acids, accession code NP-001003794 and short isoform 2; 303 amino acids, accession code NP-009214) (Labar et al., 2010). The findings presented here were obtained using the long isoform (hMAGL-long). It is not currently known why different MAGL isoforms exist or whether they display distinct catalytic properties and/or substrate preferences. To unveil possible differences in substrate preferences, we transiently expressed the human and mouse short MAGL isoforms (hMAGL-short and mMAGL-short, respectively) in HEK293 cells and visualized the SDS-PAGE mobility pattern of catalytically active isoforms using ABPP with the active site serine-targeting probe TAMRA-FP. As

527

528

Savinainen et al.

Molecular Modeling Highlights Key Interactions of PG-G Species with the MAGL Active Site. We considered various possibilities to explain why 15d-PGJ2-G was the preferred MAGL substrate in the PG-G series. The findings herein indicated that substrate availability might be one determinant. Moreover, it should be noted that 15d-PGJ2-G represented the 2-isomer, while the other PG-G species were 1(3)-isomers. It is known from previous studies that MAGL hydrolyses 1(3)- and 2-monoacylglycerols with similar rates (Saario et al., 2004; Navia-Paldanius et al., 2012; Laitinen et al., 2014). Unfortunately, restricted availability of additional PG-G species precluded further studies that would directly compare utilization of PG-G isomer pairs. To gain molecular level insights into this issue, molecular modeling studies were employed. These studies suggested that the PG-G species all converged well to the MAGL active site regardless of the isomer; we found no major differences in the docking of the glycerol moiety. However, the prostaglandin moiety in 15d-PGJ2-G offered less hydrophilic interactions with the MAGL residues as compared with the other PG-G species (Fig. 5). Interestingly, the additional hydrogen bond contacts of

Fig. 4. Substrate availability for hMAGL-catalyzed hydrolysis in assays with or without BSA. (A) Michaelis-Menten kinetics for hMAGL-catalyzed hydrolysis of 15d-PGJ2-G under routine assay conditions containing 0.5% BSA (w/v). To ensure activity measurements within tolerable substrate consumption and the linear range of glycerol assay, hMAGL-HEK lysates were diluted 1:10 or 1:20 with Mock-HEK lysate. Lysates (0.3 mg/well) were incubated with the indicated concentrations of 15d-PGJ2-G, and the Km and Vmax values were determined at the time point of 30 minutes where substrate availability was not rate limiting. (B) The Km and Vmax values were calculated as nonlinear regressions using GraphPad Prism 5.0 for Windows. The lines represent Lineweaver-Burk plots of the data in A. Values are mean 6 S.D. from two independent experiments. (C) Substrate availability for MAGL-catalyzed hydrolysis in incubation with (+) or without (2) BSA. Lysates (0.3 mg/well) of HEK293 cells transiently overexpressing hMAGL were incubated together with the indicated substrates (25 mM final concentration) in the presence or absence of 0.5% BSA. Glycerol output was determined at the time point of 30 minutes. Data are mean + S.E.M. of duplicate wells from two independent experiments. Statistical comparisons were done using one-way analysis of variance, followed by Tukey’s multiple comparisons, and statistical significance is indicated with asterisks (*P , 0.05; **P , 0.01). (D) Michaelis-Menten kinetics for hMAGL-catalyzed hydrolysis of selected substrates (2-AG, 15d-PGJ2-G, 1-DG and PGD2-G) in incubations without BSA. Cellular lysates (0.3 mg/well) were incubated with the indicated substrate concentrations, and glycerol output was determined at the time point of 30 minutes where substrate availability was not rate limiting. The Km values were calculated as nonlinear regressions using GraphPad Prism 5.0 for Windows and were 9.7 6 3.6, 12.2 6 4.6, and 20.0 6 2.6 mM for 2-AG, 15d-PGJ2-G, and 1-DG, respectively. Due to negligible hydrolysis, no accurate Km value could be determined for PGD2-G. Values are mean 6 S.E.M. from three to four independent experiments except for PGD2-G where n = 2.

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015

values for 15d-PGJ2-G were similar (12–17 mM) regardless of the enzyme dilution (Fig. 4, A and B). For the enzyme dilutions 1:20 and 1:10, the maximal velocities (Vmax) were 38 6 2 and 75 6 5 nmol/mg per minute, respectively. We were curious to learn whether omission of BSA might affect substrate availability, so we compared the use of selected substrates (2-AG, 1-DG, and 15d-PGJ2-G) for MAGL-catalyzed hydrolysis in incubations with or without BSA. The substrates were tested at 25 mM final concentrations with the outcome that while the hydrolysis rates of 15d-PGJ2-G were comparable under both conditions, those of 2-AG were significantly higher and those of 1-DG significantly lower in incubations without BSA (Fig. 4C). Using Michaelis-Menten kinetics, we determined the Km values for hMAGL-catalyzed hydrolysis of 2-AG, 15d-PGJ2-G, 1-DG, and PGD2-G in incubations without BSA. Under such conditions, the Km values were 9.7 6 3.6, 12.2 6 4.6, and 20.0 6 2.6 mM for 2-AG, 15d-PGJ2-G, and 1-DG, respectively (Fig. 4D). Noteworthy, hardly any hydrolysis of PGD2-G was detected in the absence of BSA, precluding accurate Km determinations (Fig. 4D).

Prostaglandin Glycerol Esters as In Vitro Substrates of MAGL

center was hydrolyzed at higher rates as compared with the other tested PG-G species. Activation of Nrf2 Signaling Pathway; Both 15d-PGJ2 and 15d-PGJ2-G stimulate nuclear translocation of Nrf2 in HUVEC endogenously expressing MAGL. Previous studies have shown that electrophilic compounds such as 15d-PGJ2 can signal via several pathways, including Nrf2 and peroxisome proliferator-activated receptor-g (PPARg) mediated pathways, to evoke cellular responses (Strauss and Glass 2001; Kansanen et al., 2009, 2012; Harmon et al., 2011; HaskewLayton et al., 2013). As 15d-PGJ2 and 15d-PGJ2-G possess two highly electrophilic a,b-unsaturated carbonyls, both compounds could engage similar signaling functions. Whether 15d-PGJ2-G is capable of mediating such functions is not yet known, although recent studies have implicated that 15d-PGJ2-G may act as a PPARg activating ligand, capable of suppressing nuclear factor of activated T cells (NFAT) transcriptional activity and interleukin-2 (IL-2) expression in activated T cells (Raman et al., 2011, 2012). However, time-dependent hydrolysis of 15d-PGJ2-G to 15d-PGJ2 likely complicates straightforward conclusions that could be drawn from studies requiring prolonged incubation times for a functional readout. To address this issue, we used HUVEC, which have shown to be a useful cellular model in studies exploring electrophile signaling (Jyrkkänen et al.,

Fig. 5. Docking of the studied PG-G species into the MAGL active site. Selected docking poses of (A) 15d-PGJ2-G (green), (B) PGD2-G (orange), (C) PGE2-G (magenta), and PGF2a-G (cyan) to the MAGL crystal structure (pdb:3PE6) illumining the key hydrogen bonding interaction of favorable docking poses. Catalytic triad residues (Ser122-Asp239-His269) are colored using yellow carbons, and the residues forming the oxyanion hole (Met123 and Ala51) are also illustrated. Selected substrate specific hydrogen bond forming residues are also shown. In the case of the PGD2-G (B), 1(3)-glycerol moiety is in S-configuration, whereas panels (C) and (D) represent the R-isomers of PGE2-G and PGF2a-G, respectively.

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015

the other PG-G species were located to the lid area at the active site entry (Supplemental Fig. 2). Theoretically, such contacts may retard the release of the prostaglandin products, potentially slowing down hydrolysis rates of the 1(3)-isomers. It is also worth noticing that the 1(3)-isomers possess a chiral carbon, indicating that they represent a racemic mixture of 1-G(R) and 1-G(S) forms. Docking the R and S forms individually to the MAGL active site gave practically similar hydrogen bond interactions with the key active site residues Met123, Tyr194, Ser122, Ala51, and His121 (Fig. 5, B–D). In the case of 15d-PGJ2-G, although contacts to Met123, Ser122, Ala51, and His121 were well preserved, hydrogen bonding to Tyr194 was not always present (Fig. 5A). Collectively, these observations suggest that the chiral and flexible 1(3)-PG-G isomers are able to form a set of similar hydrogen bond interactions in various orientations with active site residues Tyr194, Ser122, and His121, in accordance with earlier findings concerning MAG substrates (Bertrand et al., 2010; Laitinen et al., 2014). The importance of Tyr194 was highlighted in our previous study (Laitinen et al., 2014) where mutation of this residue was shown to affect equally hydrolysis rates of 2- and 1(3)-MAG species, while mutation of Cys242 disrupted the balanced hydrolysis of the isomers in favor of the 2-isomers. These findings could at least partly explain why 15d-PGJ2-G without the chiral

529

530

Savinainen et al.

2008; Kansanen et al., 2009). Competitive ABPP revealed that HUVEC endogenously express MAGL, as evidenced by the presence of a JJKK-048–sensitive protein doublet migrating at the 33–35 kDa range in SDS-PAGE (Fig. 6A). JJKK-048 is a highly potent MAGL-selective inhibitor (Aaltonen et al., 2013). Hydrolase assays indicated that MAGL fully accounted for 15d-PGJ2-G hydrolysis in HUVEC lysates, as this activity was potently (IC50 ∼0.2 nM) and comprehensively blocked by JJKK-048 (Fig. 6B). Using intact HUVEC, we compared the capacity of 15d-PGJ2 and 15d-PGJ2-G to induce nuclear enrichment of Nrf2, an early measurable step in Nrf2 signal transduction (Fig. 6C). Our previous studies have shown that Nrf2 translocation is detectable in HUVEC after 2-hour stimulation with oxidized phospholipids (Jyrkkänen et al., 2008). To achieve optimal signal and to avoid prolonged incubation times, we chose the

2-hour time point for closer inspection. The results indicated that when tested at 5 mM concentration 15d-PGJ2 and 15d-PGJ2-G to the same extent evoked nuclear enrichment of Nrf2 in the endothelial cells. Of note, the MAGL-selective inhibitor JJKK-048 had no effect on this response (Fig. 6D). Thus, in HUVEC with endogenous MAGL expression, the electrophiles similarly induced nuclear translocation of Nrf2, and this response was independent of MAGL activity. 15d-PGJ2-G Hydrolysis in Intact HEK293 Cells. We assessed 15d-PGJ2-G hydrolysis in situ using HEK293 cells with or without stable 3HA-hMAGL expression (Savinainen et al., 2010). Cells were pretreated for 1 hour with solvent (DMSO or the MAGL-inhibitor JJKK-048) and incubated thereafter for further 2 hours in HBS with or without 15d-PGJ2-G (25 mM final concentration). The buffer was sampled at various time points, and the amount of liberated glycerol was determined. The results

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015 Fig. 6. 15d-PGJ2 and 15d-PGJ2-G both stimulate nuclear translocation of Nrf2 in HUVEC endogenously expressing MAGL. (A) Competitive APBB of HUVEC lysates unveiling endogenous expression of catalytically active serine hydrolases, including MAGL (a protein doublet migrating at 33–35 kDa whose labeling by the active-site directed probe is prevented by the MAGL-selective inhibitor JJKK-048 (100 nM). The image is representative from three separate ABPP runs with a similar outcome (B). MAGL fully accounts for 15d-PGJ2-G hydrolysis in HUVEC lysates, as evidenced by comprehensive blockade of this hydrolysis by the MAGL-selective inhibitor JJKK-048 (IC50 178 pM) and by the fact that JJKK-048 inhibits 15d-PGJ2-G hydrolysis to the same extent as seen with the globally acting serine hydrolase inhibitor methylarachidonoyl fluorophosphonate (MAFP) (1024 M). (C) Schematics and temporal aspects of stress-activated Nrf2 signaling pathway. Under basal conditions, Nrf2 is targeted for proteosomal degradation. Electrophilic or oxidative stress leads to Nrf2 translocation into the nucleus to increase transcription of target genes such as heme oxygenase 1 (HMOX1), glutamate-cysteine ligase modifier subunit (GCLM), and NAD(P)H quinone oxidoreductase-1 (NQO1), all containing the antioxidant response element (ARE) in the 59-regulatory region. (D) 15d-PGJ2 and 15d-PGJ2-G stimulate nuclear translocation of Nrf2 in HUVEC equally well. The MAGL inhibitor JJKK-048 (2  1027 M) has no effect on this response. Bottom: representative Western blots of the nuclear Nrf2 signal in comparison with lamin B1 that was used as a loading control. In (B), the data are mean 6 S.E.M. of duplicate wells from two independent experiments. In (D), the data are mean + S.E.M. from three independent experiments. Statistical comparisons were done using one-way analysis of variance, followed by Tukey’s multiple comparisons, and statistical significance is indicated with asterisks (***P , 0.001).

Prostaglandin Glycerol Esters as In Vitro Substrates of MAGL

531

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015

indicated that significant 15d-PGJ2-G hydrolysis was evident in Mock cells only after 60 minutes, whereas in 3HA-hMAGL-HEK293 cells the substrate was efficiently hydrolyzed already at the earliest time point examined (Fig. 7). As expected, 15d-PGJ2-G hydrolysis in hMAGL-HEK293 cells was sensitive to JJKK-048, whereas this inhibitor had no effect on 15d-PGJ2-G hydrolysis in Mock-cells. Collectively, these experiments revealed relatively low endogenous hydrolytic activity toward 15d-PGJ2-G in HEK293 cells, whereas with hMAGL expression, 15d-PGJ2-G hydrolysis was clearly accelerated. We next wished to systematically compare the ability of 15d-PGJ2 and 15d-PGJ2-G to activate the Nrf2 signaling pathway, and we used multiple readouts for Nrf2 activation, including nuclear localization, luciferase reporter assays, and the ability of the electrophiles to set in train Nrf2 target gene expression. 15d-PGJ2 and 15d-PGJ2-G Similarly Activate Nrf2 Signaling Pathway and Its Target Gene Expression in HEK293 Cells with or without hMAGL Overexpression. As in the studies with HUVEC, we used Western blot analysis to monitor nuclear enrichment of Nrf2 after 2-hour stimulation with 15d-PGJ2 or 15d-PGJ2-G (5 mM final concentration) (Fig. 8A). The outcome was similar in cells with or without 3HA-hMAGL expression, although a nonsignificant trend toward potentiation of the 15d-PGJ2-G response in 3HA-hMAGL-HEK cells treated with the MAGL-inhibitor JJKK-048 was evident. Doseresponse studies of electrophile-driven antioxidant response element (ARE) activation revealed that 15d-PGJ2-G stimulated this response significantly better as compared with 15d-PGJ2 (Fig. 8B). Again, the outcome was similar regardless of 3HA-hMAGL expression. As a distal step in Nrf2 signaling, induction of three Nrf2 target genes (HMOX1, GCLM, and NQO1) was examined in response to electrophile stimulation (0–5 mM). These results indicated that 15d-PGJ2-G induced HMOX1 mRNA levels significantly more efficiently than did 15d-PGJ2 (Fig. 8C). A similar, albeit nonsignificant, trend was also evident for the induction of GCLM and NQO1 (Fig. 8, D and E). The observed responses were similar between Mock-HEK and 3HA-hMAGL-HEK cells, the sole exception being that no statistically significant increase in NQO1 transcripts was detected in 3HA-hMAGLHEK cells (Fig. 8, C–E).

Discussion In this study, we have demonstrated that the human endocannabinoid hydrolases MAGL, ABHD6, and ABHD12 can variably hydrolyze PG-G species in vitro and that hMAGL robustly hydrolyzed all tested PG-G species, and at rates rivaling those of the best MAG substrates evaluated under comparable conditions. We demonstrated that hMAGL showed particular preference toward 15d-PGJ2-G. Michaelis-Menten kinetics revealed that hMAGL avidly (Km ∼15 mM) bound 15d-PGJ2-G. Whereas the availability of 15d-PGJ2-G for MAGL-catalyzed hydrolysis was similar from the buffer with or without BSA, the importance of the protein carrier in delivery of PGD2-G for enzymatic hydrolysis was demonstrated. Molecular docking studies provided atomic level insights into variable interactions of the PG-G isomers with the MAGL active site. We demonstrated that the MAGL isoforms shared a similar substrate profile, and all robustly hydrolyzed 15d-PGJ2-G. We showed that hMAGL used 15d-PGJ2-G also in living HEK293

Fig. 7. hMAGL hydrolyses 15d-PGJ2-G in living HEK293 cells. HEK293 cells with (gray columns) or without (clear columns) stable overexpression of 3HA-hMAGL were seeded on a 96-well plate; after 2 days, the cells were switched to HBS and pretreated for 1 hour with dimethylsulfoxide or the MAGL-selective inhibitor JJKK-048 at the indicated concentrations; after which, fresh HBS with or without 15d-PGJ2-G (25 mM final concentration) was added, and the cells were incubated further for 120 minutes at 37°C. Aliquots (40 ml) of buffer were drawn at the time points 0, 30, 60, and 120 minutes after substrate addition, and the glycerol content was determined based on a glycerol standard. Data are mean + S.E.M. of triplicate wells from a representative experiment that was repeated twice with similar outcome. Statistical comparisons were done using one-way analysis of variance, followed by Tukey’s multiple comparisons, and statistical significance is indicated with asterisks (*P , 0.05; ***P , 0.001).

cells engineered to express this enzyme. In HUVEC lysates, endogenous MAGL fully accounted for 15d-PGJ2-G hydrolysis. To explore the consequences of this activity, we assessed the

532

Savinainen et al.

ability of 15d-PGJ2-G to activate the canonical Nrf2 signaling pathway known to be used by the hydrolysis end product 15d-PGJ2, and these studies unveiled for the first time that 15d-PGJ2 and 15d-PGJ2-G both activated the Nrf2 pathway as well as transcription of several target genes of this pathway. Our findings challenge previous claims regarding the ability of MAGL to use PG-G species and extend the MAGL substrate profile beyond the classic MAG species. MAGL Acts as a PG-G Hydrolase In Vitro; BSA Variably Affects Substrate Availability Previously, the ability of PGD2-G, PGE2-G, and PGF2a-G to inhibit 2-oleoylglycerol (2-OG) hydrolysis by rat brain cytosolic MAGL preparation was examined (Fowler and Tiger

2005). This study indicated that the tested PG-G species were weak inhibitors of 2-oleoylglycerol hydrolysis, suggesting that these compounds were not efficiently competing with the substrate. On the other hand, 15d-PGJ2 at the low-micromolar range (IC50 ∼50 mM) was reported to inhibit hMAGL-mediated 4-nitrophenylacetate hydrolysis (Björklund et al., 2010). In light of the present findings, this could represent end-product inhibition of MAGL activity. We tested this possibility by using a competitive approach where MAGL-catalyzed hydrolysis of 15d-PGJ2-G and 1-AG was evaluated in the presence of increasing concentrations of 15d-PGJ2. We found, however, that the end product did not affect the hydrolysis rates of these substrates, not even when present at 8-fold excess quantities (Supplemental Fig. 3).

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015

Fig. 8. Comparative analysis of activation of Nrf2 signaling pathway by 15d-PGJ2 and 15d-PGJ2-G in HEK293 cells with or without hMAGL overexpression. (A) Nuclear translocation of Nrf2 in response to 5 mM 15d-PGJ2 and 15d-PGJ2-G in Mock-transfected (HEK) and hMAGL-HEK cells. Note the weak response to 15d-PGJ2 in HEK293 cells as compared with HUVEC (Fig. 6). Note also that although one-way analysis of variance indicated no statistically significant differences between the treatments, the MAGL inhibitor JJKK-048 (0.2 mM) shows a clear trend in amplifying the 15d-PGJ2-G response in hMAGL-HEK cells. Bottom: representative Western blots of the nuclear Nrf2 signal in comparison with the lamin B1 that was used as a loading control. (B) Dose-dependent antioxidant response element (ARE) activation by 15d-PGJ2 and 15d-PGJ2-G in HEK293 cells (left) and hMAGL-HEK cells (right). Dose-dependent and similar induction of transcripts of the Nrf2 target genes heme oxygenase 1 (HMOX1) (C), glutamate-cysteine ligase modifier subunit (GCLM) (D), and NAD(P)H quinone oxidoreductase-1 (NQO1) (E) by 15d-PGJ2 and 15d-PGJ2-G in HEK293 cells with or without hMAGL overexpression. Data are mean + S.E.M. from triplicate wells from a representative experiment that was repeated once with a similar outcome. Statistical comparisons were performed using one-way analysis of variance, followed by Tukey’s multiple comparisons, and statistical significance is indicated with asterisks (*P , 0.05; **P , 0.01; ***P , 0.001) or hashtags (#P , 0.05; ###P , 0.001).

Prostaglandin Glycerol Esters as In Vitro Substrates of MAGL

15d-PGJ2-G Hydrolysis Generates the Anti-Inflammatory Lipid 15d-PGJ2 Cyclopentenone prostaglandins possess highly electrophilic a,b-unsaturated carbonyls that can react with cysteine thiols,

either on glutathione or cellular proteins (Buczynski et al., 2009; Kansanen et al., 2009). In fact, 15d-PGJ2 possesses two such electrophilic centers and is considered as an endogenous anti-inflammatory lipid with a wide spectrum of physiologic activities (Scher and Pillinger 2005; Buczynski et al., 2009). For inflammation, 15d-PGJ2 plays a role in its resolution (Rajakariar et al., 2007), and it has been identified as a highaffinity activator of PPARg and Nrf2 signaling pathways (Buczynski et al., 2009; Kansanen et al., 2009). Using various readouts to monitor activation of the Nrf2 signaling pathway, we demonstrated that in HUVEC endogenously expressing MAGL, 15d-PGJ2 and 15d-PGJ2-G both increased nuclear content of Nrf2 to the same extent, and pharmacologic MAGL inhibition had no effect on this response. In HEK293 cells with or without MAGL expression, both electrophiles stimulated Nrf2 nuclear accumulation, whereas comparative doseresponse studies indicated that on a molar basis 15d-PGJ2-G was a stronger antioxidant response element (ARE) activator and inducer of Nrf2 target gene expression compared with 15d-PGJ2. As far as we are aware, this is the first study to document that a PG-G is capable of activating the Nrf2 signaling pathway, a property previously attributed to 15d-PGJ2. The signaling functions of the PG-G species are currently poorly understood, but 15d-PGJ2-G has been shown to inhibit interleukin-2 expression via PPARg (Raman et al., 2011), and PGD2-G has anti-inflammatory properties in macrophages via unknown receptor(s) (Alhouayek et al., 2013). These results combined with ours imply that COX-2–derived PG-G species have potent signaling capabilities that may be modified by PG-G hydrolytic enzymes such as MAGL and ABHD6. Possible Implications and Relevance of Enzymatic PG-G Hydrolysis The lipid signaling pathways are closely intertwined (Kukkonen 2011; Mulvihill and Nomura 2013). In the central nervous system, 2-AG acts as a lipid mediator and its enzymatic hydrolysis not only terminates endocannabinoid signaling but concomitantly feeds the key substrate for COX-2–catalyzed production of prostaglandins, a series of lipid mediators that play crucial roles in a variety of biologic responses (Hermanson et al., 2014). Furthermore, COX-2– mediated oxygenation is emerging as the third major metabolic pathway for the endocannabinoids with potential therapeutic implications (Rouzer and Marnett 2011; Hermanson et al., 2014). Under specific conditions, 2-AG–derived PG-G species are formed by COX-2 action in vitro and in vivo, and these metabolites likely have bioactivities of their own, mediated via as-yet-unidentified non-PG receptors (Rouzer and Marnett 2011; Hermanson et al., 2014). Thus, MAGL-dependent PG-G hydrolysis, if proven relevant in vivo, could be envisioned to regulate the postulated signaling functions of PG-G species. Indeed, the first in vivo evidence supporting this possibility was recently reported (Shimizu et al., 2014). We found that PGD 2-G was the preferred PG-G substrate for hABHD6, with hydrolysis rates ∼60% of those observed for 1-AG. However, this in vitro finding does not necessarily mean that ABHD6 would use this substrate in vivo. Nevertheless, this result could be relevant in light of recent findings demonstrating that pharmacologic inhibition of ABHD6 in lipopolysaccharide-induced murine macrophages led to 2-AG accumulation and channeling to COX-2–derived anti-inflammatory PG-G species, including PGD2-G, a precursor of 15d-PGJ2-G

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015

Very few studies have directly assessed the ability of MAGL to use the PG-G species. The first study (Vila et al., 2007), using purified rat MAGL and PGE2-G and PGF2a-G as the substrates, concluded that in comparison with 2-AG the PG-G species were poor MAGL substrates, a view that prevails in the literature (Rouzer and Marnett, 2011). More recently and quite the opposite, Xie et al. (2010) reported that hMAGL efficiently hydrolyzed PGE2-G and PGF2a-G. In line with these results, we found that 3HA-hMAGL readily used various PG-G species (Laitinen et al., 2014). Our present work not only confirms but also substantially extends these preliminary findings. We demonstrated that restricted substrate availability due to absence of protein carrier was a critical factor compromising successful detection of MAGL-catalyzed PGD2-G hydrolysis. This finding could potentially also explain the negative outcome of the original study (Vila et al., 2007). In contrast to the methodology used in that study, our assays as well as those used by Xie et al. (2010) contained BSA as a carrier for lipophilic substrates. We argue that inclusion of BSA is fully justified, clearly facilitating availability of substrates like 1-DG for enzymatic hydrolysis. On the other hand, the carrier may avidly bind other substrates (as illustrated here for 2-AG), restricting thereby the availability for enzymatic hydrolysis. We noted that availability of 2-AG was similarly affected by bovine and human albumin (0.5%) whereas 0.05% albumin behaved like the aqueous buffer alone (Supplemental Fig. 4). However, tight binding to BSA does not totally occlude substrate availability but is typically reflected as lowered affinity toward the enzyme (i.e., higher Km values). In the case of 2-AG, Km values around 100–200 mM were obtained for MAGL-catalyzed hydrolysis in assays containing BSA (Navia-Paldanius et al., 2012; Laitinen et al., 2014), whereas a Km value of ∼10 mM was determined here in incubations without BSA (Fig. 4D). Yet for other substrates such as 15d-PGJ2-G, hydrolysis rates (and Km values) were comparable in the presence and absence of the carrier. It is noteworthy that the robust hydrolysis of 2-AG as opposed to the barely detectable hydrolysis of PGD2-G in our assays without BSA is reminiscent of the findings of the original study with the negative outcome (Vila et al., 2007). Molecular modeling further supported favorable interaction of the PG-G species with the hMAGL active site. These studies suggested that due to lack of chiral carbonyl center, 15d-PGJ2-G (2-isomer) was more favorably posed to the MAGL active site as compared with the 1(3)-isomers possessing the chiral center. Modeling studies indicated further that in contrast with the other PG-G species, 15d-PGJ2-G possessing the most lipophilic prostaglandin tail did not form efficient hydrogen bond contacts within the lid area of MAGL. Theoretically, such contacts could retard the release of the prostaglandin products, potentially slowing down hydrolysis rates of the 1(3)-isomers. This line of reasoning is consistent with previous observations suggesting that the functionality of hMAGL includes movement of the flexible lid loop between open and closed conformations upon binding of the substrate (Karageorgos et al., 2010; 2013).

533

534

Savinainen et al.

(Alhouayek et al., 2013). Furthermore, in vivo administration of PGD2-G was shown to reduce lipopolysaccharide-induced inflammation, confirming the biologic relevance of this 2-AG metabolite. In light of our findings, ABHD6 inhibitors might not only funnel accumulating 2-AG toward PGD2-G generation, but could also concomitantly prolong its biologic half-life by preventing its breakdown. This should be a viable hypothesis to be tested in future in vivo studies. Acknowledgments

The authors thank Satu Marttila, Taija Hukkanen, Minna Glad, and Arja Korhonen for their highly competent technical assistance in the laboratory, and the CSC-IT Center for Science Ltd. for the allocation of computational resources. Authorship Contributions

References Aaltonen N, Savinainen JR, Ribas CR, Rönkkö J, Kuusisto A, Korhonen J, NaviaPaldanius D, Häyrinen J, Takabe P, and Käsnänen H et al. (2013) Piperazine and piperidine triazole ureas as ultrapotent and highly selective inhibitors of monoacylglycerol lipase. Chem Biol 20:379–390. Alhouayek M, Masquelier J, Cani PD, Lambert DM, and Muccioli GG (2013) Implication of the anti-inflammatory bioactive lipid prostaglandin D2-glycerol ester in the control of macrophage activation and inflammation by ABHD6. Proc Natl Acad Sci USA 110:17558–17563. Alhouayek M, Masquelier J, and Muccioli GG (2014) Controlling 2-arachidonoylglycerol metabolism as an anti-inflammatory strategy. Drug Discov Today 19:295–304. Bertrand T, Augé F, Houtmann J, Rak A, Vallée F, Mikol V, Berne PF, Michot N, Cheuret D, and Hoornaert C et al. (2010) Structural basis for human monoglyceride lipase inhibition. J Mol Biol 396:663–673. Björklund E, Norén E, Nilsson J, and Fowler CJ (2010) Inhibition of monoacylglycerol lipase by troglitazone, N-arachidonoyl dopamine and the irreversible inhibitor JZL184: comparison of two different assays. Br J Pharmacol 161:1512–1526. Blankman JL, Simon GM, and Cravatt BF (2007) A comprehensive profile of brain enzymes that hydrolyze the endocannabinoid 2-arachidonoylglycerol. Chem Biol 14:1347–1356. Buczynski MW, Dumlao DS, and Dennis EA (2009) Thematic Review Series: Proteomics. An integrated omics analysis of eicosanoid biology. J Lipid Res 50:1015–1038. Caterina MJ (2014) TRP channel cannabinoid receptors in skin sensation, homeostasis, and inflammation. ACS Chem Neurosci 2014:10 DOI: 10.1021/cn5000919 [published ahead of print . Chanda PK, Gao Y, Mark L, Btesh J, Strassle BW, Lu P, Piesla MJ, Zhang MY, Bingham B, and Uveges A et al. (2010) Monoacylglycerol lipase activity is a critical modulator of the tone and integrity of the endocannabinoid system. Mol Pharmacol 78:996–1003. Chen R, Zhang J, Wu Y, Wang D, Feng G, Tang YP, Teng Z, and Chen C (2012) Monoacylglycerol lipase is a therapeutic target for Alzheimer’s disease. Cell Reports 2:1329–1339. Di Marzo V, Bisogno T, and De Petrocellis L (2007) Endocannabinoids and related compounds: walking back and forth between plant natural products and animal physiology. Chem Biol 14:741–756. Fowler CJ and Tiger G (2005) Cyclooxygenation of the arachidonoyl side chain of 1-arachidonoylglycerol and related compounds block their ability to prevent anandamide and 2-oleoylglycerol metabolism by rat brain in vitro. Biochem Pharmacol 69:1241–1245. Harmon GS, Lam MT, and Glass CK (2011) PPARs and lipid ligands in inflammation and metabolism. Chem Rev 111:6321–6340. Haskew-Layton RE, Payappilly JB, Xu H, Bennett SA, and Ratan RR (2013) 15-Deoxy-D12,14-prostaglandin J2 (15d-PGJ2) protects neurons from oxidative death via an Nrf2 astrocyte-specific mechanism independent of PPARg. J Neurochem 124:536–547. Hermanson DJ, Gamble-George JC, Marnett LJ, and Patel S (2014) Substrateselective COX-2 inhibition as a novel strategy for therapeutic endocannabinoid augmentation. Trends Pharmacol Sci 35:358–367. Hurttila H, Koponen JK, Kansanen E, Jyrkkänen HK, Kivelä A, Kylätie R, YläHerttuala S, and Levonen AL (2008) Oxidative stress-inducible lentiviral vectors for gene therapy. Gene Ther 15:1271–1279.

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015

Participated in research design: Savinainen, Kansanen, Lehtonen, Levonen, J.T. Laitinen. Conducted experiments: Savinainen, Kansanen, Pantsar, NaviaPaldanius, Parkkari, Lehtonen, T. Laitinen, Nevalainen, Poso, Levonen, J.T. Laitinen. Performed data analysis: Savinainen, Kansanen, Lehtonen, Pantsar, Parkkari, T. Laitinen, J.T. Laitinen. Wrote or contributed to the writing of the manuscript: Savinainen, Kansanen, Parkkari, Lehtonen, T. Laitinen, Nevalainen, Poso, Levonen, J.T. Laitinen.

Jyrkkänen HK, Kansanen E, Inkala M, Kivelä AM, Hurttila H, Heinonen SE, Goldsteins G, Jauhiainen S, Tiainen S, and Makkonen H et al. (2008) Nrf2 regulates antioxidant gene expression evoked by oxidized phospholipids in endothelial cells and murine arteries in vivo. Circ Res 103:e1–e9. Kano M, Ohno-Shosaku T, Hashimotodani Y, Uchigashima M, and Watanabe M (2009) Endocannabinoid-mediated control of synaptic transmission. Physiol Rev 89:309–380. Kansanen E, Jyrkkänen HK, and Levonen AL (2012) Activation of stress signaling pathways by electrophilic oxidized and nitrated lipids. Free Radic Biol Med 52:973–982. Kansanen E, Kivelä AM, and Levonen AL (2009) Regulation of Nrf2-dependent gene expression by 15-deoxy-Delta12,14-prostaglandin J2. Free Radic Biol Med 47:1310– 1317. Karageorgos I, Tyukhtenko S, Zvonok N, Janero DR, Sallum C, and Makriyannis A (2010) Identification by nuclear magnetic resonance spectroscopy of an active-site hydrogen-bond network in human monoacylglycerol lipase (hMGL): implications for hMGL dynamics, pharmacological inhibition, and catalytic mechanism. Mol Biosyst 6:1381–1388. Karageorgos I, Wales TE, Janero DR, Zvonok N, Vemuri VK, Engen JR, Makriyannis A, and Makriyannis A (2013) Active-site inhibitors modulate the dynamic properties of human monoacylglycerol lipase: a hydrogen exchange mass spectrometry study. Biochemistry 52:5016–5026. Karlsson M, Reue K, Xia YR, Lusis AJ, Langin D, Tornqvist H, and Holm C (2001) Exon-intron organization and chromosomal localization of the mouse monoglyceride lipase gene. Gene 272:11–18. Kukkonen JP (2011) A ménage à trois made in heaven: G-protein-coupled receptors, lipids and TRP channels. Cell Calcium 50:9–26. Labar G, Wouters J, and Lambert DM (2010) A review on the monoacylglycerol lipase: at the interface between fat and endocannabinoid signalling. Curr Med Chem 17:2588–2607. Laitinen T, Navia-Paldanius D, Rytilahti R, Marjamaa JJ, Karízková J, Parkkari T, Pantsar T, Poso A, Laitinen JT, and Savinainen JR (2014) Mutation of Cys242 of human monoacylglycerol lipase disrupts balanced hydrolysis of 1- and 2-monoacylglycerols and selectively impairs inhibitor potency. Mol Pharmacol 85:510–519. Levonen AL, Dickinson DA, Moellering DR, Mulcahy RT, Forman HJ, and DarleyUsmar VM (2001) Biphasic effects of 15-deoxy-delta(12,14)-prostaglandin J(2) on glutathione induction and apoptosis in human endothelial cells. Arterioscler Thromb Vasc Biol 21:1846–1851. Mulvihill MM and Nomura DK (2013) Therapeutic potential of monoacylglycerol lipase inhibitors. Life Sci 92:492–497. Navia-Paldanius D, Savinainen JR, and Laitinen JT (2012) Biochemical and pharmacological characterization of human a/b-hydrolase domain containing 6 (ABHD6) and 12 (ABHD12). J Lipid Res 53:2413–2424. Nomura DK, Morrison BE, Blankman JL, Long JZ, Kinsey SG, Marcondes MC, Ward AM, Hahn YK, Lichtman AH, and Conti B et al. (2011) Endocannabinoid hydrolysis generates brain prostaglandins that promote neuroinflammation. Science 334:809–813. Piro JR, Benjamin DI, Duerr JM, Pi Y, Gonzales C, Wood KM, Schwartz JW, Nomura DK, and Samad TA (2012) A dysregulated endocannabinoid-eicosanoid network supports pathogenesis in a mouse model of Alzheimer’s disease. Cell Reports 1:617–623. Rajakariar R, Hilliard M, Lawrence T, Trivedi S, Colville-Nash P, Bellingan G, Fitzgerald D, Yaqoob MM, and Gilroy DW (2007) Hematopoietic prostaglandin D2 synthase controls the onset and resolution of acute inflammation through PGD 2 and 15-deoxyD12,14 PGJ 2 . Proc Natl Acad Sci USA 104:20979– 20984. Raman P, Kaplan BL, and Kaminski NE (2012) 15-Deoxy-D12,14-prostaglandin J2-glycerol, a putative metabolite of 2-arachidonyl glycerol and a peroxisome proliferator-activated receptor g ligand, modulates nuclear factor of activated T cells. J Pharmacol Exp Ther 342:816–826. Raman P, Kaplan BL, Thompson JT, Vanden Heuvel JP, and Kaminski NE (2011) 15-Deoxy-D12,14-prostaglandin J2-glycerol ester, a putative metabolite of 2-arachidonyl glycerol, activates peroxisome proliferator activated receptor gamma. Mol Pharmacol 80:201–209. Rouzer CA and Marnett LJ (2011) Endocannabinoid oxygenation by cyclooxygenases, lipoxygenases, and cytochromes P450: cross-talk between the eicosanoid and endocannabinoid signaling pathways. Chem Rev 111:5899–5921. Saario SM, Savinainen JR, Laitinen JT, Järvinen T, and Niemi R (2004) Monoglyceride lipase-like enzymatic activity is responsible for hydrolysis of 2-arachidonoylglycerol in rat cerebellar membranes. Biochem Pharmacol 67:1381–1387. Savinainen JR, Saario SM, and Laitinen JT (2012) The serine hydrolases MAGL, ABHD6 and ABHD12 as guardians of 2-arachidonoylglycerol signalling through cannabinoid receptors. Acta Physiol (Oxf) 204:267–276. Savinainen JR, Saario SM, Niemi R, Järvinen T, and Laitinen JT (2003) An optimized approach to study endocannabinoid signaling: evidence against constitutive activity of rat brain adenosine A1 and cannabinoid CB1 receptors. Br J Pharmacol 140:1451–1459. Savinainen JR, Yoshino M, Minkkilä A, Nevalainen T, and Laitinen JT (2010) Characterization of binding properties of monoglyceride lipase inhibitors by a versatile fluorescence-based technique. Anal Biochem 399:132–134. Schalk-Hihi C, Schubert C, Alexander R, Bayoumy S, Clemente JC, Deckman I, DesJarlais RL, Dzordzorme KC, Flores CM, and Grasberger B et al. (2011) Crystal structure of a soluble form of human monoglyceride lipase in complex with an inhibitor at 1.35 Å resolution. Protein Sci 20:670–683. Scher JU and Pillinger MH (2005) 15d-PGJ2: the anti-inflammatory prostaglandin? Clin Immunol 114:100–109. Schlosburg JE, Blankman JL, Long JZ, Nomura DK, Pan B, Kinsey SG, Nguyen PT, Ramesh D, Booker L, and Burston JJ et al. (2010) Chronic monoacylglycerol lipase

Prostaglandin Glycerol Esters as In Vitro Substrates of MAGL blockade causes functional antagonism of the endocannabinoid system. Nat Neurosci 13:1113–1119. Shimizu T, Tanaka K, Nakamura K, Taniuchi K, Yawata T, Higashi Y, Ueba T, Dimitriadis F, Shimizu S, and Yokotani K et al. (2014) Possible involvement of brain prostaglandin E2 and prostanoid EP3 receptors in prostaglandin E2 glycerol ester-induced activation of central sympathetic outflow in the rat. Neuropharmacology 82:19–27. Straus DS and Glass CK (2001) Cyclopentenone prostaglandins: new insights on biological activities and cellular targets. Med Res Rev 21:185–210. Vila A, Rosengarth A, Piomelli D, Cravatt B, and Marnett LJ (2007) Hydrolysis of prostaglandin glycerol esters by the endocannabinoid-hydrolyzing enzymes,

535

monoacylglycerol lipase and fatty acid amide hydrolase. Biochemistry 46:9578– 9585. Xie S, Borazjani A, Hatfield MJ, Edwards CC, Potter PM, and Ross MK (2010) Inactivation of lipid glyceryl ester metabolism in human THP1 monocytes/macrophages by activated organophosphorus insecticides: role of carboxylesterases 1 and 2. Chem Res Toxicol 23:1890–1904.

Address correspondence to: Jarmo T. Laitinen, Institute of Biomedicine, School of Medicine, University of Eastern Finland, Kuopio campus, POB 1627, 70211 Kuopio, Finland. E-mail: [email protected]

Downloaded from molpharm.aspetjournals.org at ASPET Journals on October 22, 2015